Available online at www.sciencedirect.

com

ScienceDirect
Energy Procedia 63 (2014) 1835 – 1841

GHGT-12

CO2 Absorption into Aqueous Amine Blends containing

Benzylamine (BZA), Monoethanolamine (MEA), and Sterically
Hindered / Tertiary Amines
William Conway*, Yaser Beyad, Paul Feron, Gilles Richner, and Graeme Puxty
CSIRO Energy Flagship, Mayfield West NSW 2304 Australia

Abstract

To date, a single mono-amine solvent with ideal properties for CO2 capture processes is yet to be found. As a compromise,
solvent formulations incorporating two or more individual amines into a single solvent blend have been proposed and are gaining
momentum as next generation capture solvents. Recently we have identified benzylamine (BZA) for use as a CO2 capture
solvent. A significant drawback with BZA is the formation of stable carbamate salts at high CO2 loadings which has severely
limited the concentration of BZA which can be used. Thus, BZA is ideally placed as a candidate for use in formulations with
other amines.
In this study, we investigate CO2 absorption into aqueous solvent blends containing BZA, MEA, sterically hindered, and
tertiary amines using a wetted-wall column contactor at 40oC, and simulations of the vapour liquid equilibrium behaviour of the
blends using a software tool developed in Matlab. Overall CO2 mass transfer coefficients, KG, at 40oC are reported for blends
containing 3.0M BZA together with 3.0M of the second amine component. Several of the formulations were found to possess
significantly faster absorption rates than in standalone MEA solutions at similar total amine concentrations, and CO2 loadings, as
in the blends. A formulation containing BZA and MEA was found to display the fastest CO2 absorption rates due to the blend
containing two fast reacting components. Highly branched and sterically hindered amines in formulations with BZA limit the
CO2 absorption rate due to the parallel impact of the amine structure on the physical properties of these solutions. Absorption
capacities at 40oC are similar in each of the blends with 6.0M MEA, however cyclic capacities are significantly higher in the
majority of the BZA blends with the largest improvement observed being 92% in a blend containing 3.0M BZA and 3.0M AMP.
Overall absorption enthalpies in all of the BZA blends here were found to be larger than that observed in standalone MEA due
largely to the contribution from the larger protonation enthalpy of BZA.
© 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
© 2013 The Authors. Published by Elsevier Ltd.
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Selection and peer-review under responsibility of GHGT.
Peer-review under responsibility of the Organizing Committee of GHGT-12

* Corresponding author. Tel.: +61 2 49 606098

E-mail address: will.conway@csiro.au

1876-6102 © 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Peer-review under responsibility of the Organizing Committee of GHGT-12
doi:10.1016/j.egypro.2014.11.191
1836 William Conway et al. / Energy Procedia 63 (2014) 1835 – 1841

Keywords: Benzylamine Kinetics, Wetted Wall Contactor, Post Combustion Capture, Mass Transfer, Equilibrium Modelling, Thermodynamics

1. Introduction

The alkanolamine MEA has remained the industrial standard solvent for CO2 capture processes for several
decades due to its simple chemical structure and rapid reaction rate with CO2. Several inherent issues with MEA and
similar solvents exist including large energy demand to regenerate CO2, significant corrosion, degradation when
exposed to trace flue gas impurities (SOx, and NOx), as well as considerable oxidative and thermal degradation. The
above issues, among several others, must be addressed in order for chemical solvents to obtain and sustain
reasonable capture performance.
Significant resources are being invested into the research and development of alternative aqueous amine solvents
to address negative characteristics with traditional solvents. However, despite these efforts, little improvement has
been offered by a single amine solvent alone. The prospect for improved solvents can be achieved through the use of
formulated solvents or amine blends containing two or more individual amine components which have been
combined into a single solvent mixture. Typically, a rapid reacting amine such as MEA or piperazine (PZ) is
combined into a blend with a slower reacting, but highly efficient amine, in terms of its absorption capacity. Such
solvents are rapidly gaining momentum as the next generation of CO2 capture solvents by offering a compromise
between the rapid kinetic performance and CO2 capacity of a solvent.
Recently, we have identified and pursued benzylamine (BZA) for potential use as a CO2 capture solvent given its
similar reactivity to MEA, low corrosivity, low viscosity in the absence of CO 2, miscibility with water, and large
protonation enthalpy.1 However, the formation of stable carbamate salts at high CO 2 loadings and high BZA
concentrations limits the aqueous concentration of BZA to 3.0M before precipitation issues arise. The formulation of
BZA into a blended solvent is an approach which can potentially be used to improve the overall concentration of
amine in a solution with BZA. The use of such a blend would ideally act to combine the fast CO 2 absorption rate in
aqueous BZA solutions together with potential improvements in the equilibrium efficiency of the solvent for CO2
(absorption and cyclic capacity) offered by highly concentrated amine solutions.
In this work, we investigate CO2 absorption into aqueous blends containing BZA with a number of candidate
sterically hindered and tertiary amines in an effort to yield the best combination of mass transfer and CO 2 cyclic
capacity performance. A combination of BZA with MEA was additionally trialled to determine the effect of a blend
containing two fast reacting amines which react to form carbamate. Overall CO2 mass transfer coefficients have been
determined using a wetted-wall contactor at 40oC for each of the blends including measurements of CO2 absorption
into partially CO2 loaded solutions. Vapour liquid equilibrium behaviour in each of the BZA blends has been
estimated here using predictions of chemical speciation at relevant temperature conditions allowing for a comparison
of the blend performances.

2. Experimental Section

2.1. Chemicals

Benzylamine (BZA) monoethanolamine (MEA), 2-amino-2-methyl-1-propanol (AMP), N,N-

dimethylethanolamine (N,N-DMEA), N,N-diethylethanolamine (N,N-DEEA), 2- dimethylamino-2-methyl-1-
propanol (2,2-M-1-P), and 3-quinuclidinol (3-QUIN) were all purchased from Sigma Aldrich and used without
further purification. Chemical structures of the amines are shown in Table 1. CO2 gas (99.99%) and N2 gas (99.99%)
were purchased from coregas Australia.
 William Conway et al. / Energy Procedia 63 (2014) 1835 – 1841 1837

Table 1. Amine structures

Amine Structure
Benzylamine (BZA)

Monoethanolamine (MEA)

2-amino-2-methyl-1-propanol (AMP)

N,N-dimethylethanolamine (N,N-DMEA)

N,N-diethylethanolamine (N,N-DEEA)

2-dimethylamino-2-methyl-1-propanol (2,2-M-1-P)

3-Quinuclidinol (3-QUIN)

2.2 Sample preparation

Aqueous solutions of the BZA blends were prepared by weighing the required amount of the amine components
into analytical glassware. All solutions were made up to volume using de-ionized water. A series of solutions
covering a range of CO2 loadings were prepared by mass difference using a simple reactor system which consisted
of a sparging frit, glass reactor vessel, gas delivery system, and an analytical balance. A pure CO 2 gas stream was
introduced to the reactor system and sparged through the frit into a known volume of amine solution (600mL)
positioned on the analytical balance. Cooling water passing from a water bath was circulating through a condenser
positioned on the outlet of the reactor vessel to minimize potential mass lost from the sample due to evaporation of
water or solvent. Given the relatively large volumes of amine solution, and the mass of CO 2 delivered into each
sample, the measured mass change in each case was assumed to be equivalent to the amount of CO 2 delivered.

2.3 Wetted wall column gas-liquid contactor

The wetted-wall column contactor used in this work, and the procedure by which it was operated, was similar to
that as detailed used in our previous studies.2,3 A short description of the setup and experimental protocol is outlined
in the following section for clarity. A stainless steel wetted-wall column measuring 8.21cm in length and 1.27cm in
diameter was used here. The column was enclosed by a glass jacket which was heated by circulating water from a
Techne TU-20D water bath held at the set point temperature. A synthetic gas containing CO2 and N2 was prepared
by passing pure streams of the individual gases through Bronkhorst mass flow controllers. The concentration of CO 2
in the gas was adjusted by varying the ratio and flow rates of the individual gas components using the mass flow
controllers. A total gas flow rate of 5.0 L/min was employed for all measurements. Before entering at the base of the
column, the gas was passed through a stainless steel coiled tube and a pre-saturator submersed in the water bath
which acted to increase the gas temperature (up to the set point) and saturate the dry gas with water.
Approximately 600mL of amine solution was placed into a glass Schott bottle which acted as the solvent
1838 William Conway et al. / Energy Procedia 63 (2014) 1835 – 1841

reservoir which was submersed into a water bath to maintain the temperature of the liquid at the set point. The
amine solution was circulated from the reservoir up through the inside of the column using a Masterflex peristaltic
pump fitted with Viton rubber tubing. Once at the peak of the column the solvent exits smoothly through small holes
before falling under gravity where it was counter currently contacted with the synthetic CO 2/N2 gas introduced at the
base of the column. Total liquid flow rate in the column was monitored by a calibrated liquid flow meter and
maintained at 121.4 mL/min. The temperature of the liquid was monitored by a temperature thermocouple located in
the liquid line exiting the base of the column as the solution is returned, via the peristaltic pump, to the solvent
reservoir in the water bath.
The composition of the gas was monitored using a Horiba VA-3000 IR gas analyser. The CO2 absorption flux,
expressed in units of mmol.s-1.m-2, was determined using the concentrations of CO2 entering (inlet) and exiting
(outlet) the wetted-wall column. The former was measured while bypassing the column and measuring the gas
directly without prior exposure to the amine liquid. Overall CO2 mass transfer coefficients, KG, were determined
from linear plots of the absorption flux over a series of applied CO2 concentrations (PCO2, determined as the log
mean of the inlet and outlet CO2 partial pressures) where the slope is equal to the overall mass transfer co-efficient.3

2.3 Vapour liquid equilibrium simulations

2.3.1 Chemical model

Equilibrium behaviour in the BZA blends, and standalone MEA solution, was predicted here using a modelling
tool developed in Matlab.4 To do so, the chemical speciation in the solutions was initially predicted using chemical
equilibrium reactions and the corresponding known published constants and enthalpies for the equilibrium reactions.
The chemical model consisted of reactions describing the aqueous chemistry of CO2 and the autoprotolysis of water,
equations (1) to (4), protonation of the amine(s), equation (5), and reactions describing the formation, and
protonation, of carbamate(s), equations (6) and (7). Equations (1) to (4) are common to all amine solutions and
values for the equilibrium constants and reaction enthalpies for these reactions were subsequently fixed in the
chemical model throughout.5

CO32-  H  mo HCO3- (1)

HCO3-  H  m o H 2CO3 (2)
H 2CO3 m o CO2 (aq) (3)

OH  H m
-
o H 2O (4)

Published values for the protonation constants and corresponding reaction enthalpies for the amines, equation (5),
were taken from our previous work.1,6 It should be noted that a term for equation (5) is required in the model for
each of the amines present in the solution. Thus, in the case of blends containing two amines here, the protonation
reaction appears in the model in duplicate.

R1R2 R3 N  H  m
o R1R2 R3 NH  (5)

Published values describing the formation of carbamate for BZA and MEA were also taken from our previous
work.1,7 Similarly, in the case of the blend containing BZA and MEA, the carbamate reactions appear in the model
in duplicate.
R1 R2 NH  HCO3- m
o R1R2 NCO2- ( H 2O) (6)

R1R2 NCO  H m
-
2 o R1R2 NCO2 H (7)

2.3.1 Absorption capacity, cyclic capacity, and absorption enthalpy calculations

From the speciation data, specifically the free (dissolved) CO 2(aq) concentration, the corresponding partial
pressure of CO2 in the gas phase was calculated using Henrys constant for CO 2 solubility in water. The absorption
capacities of the blends and MEA solutions were initially determined at 40oC up to a fixed maximum CO2 partial
pressure of 15 kPa. The above calculation was repeated at 100oC and the cyclic capacity calculated from the
 William Conway et al. / Energy Procedia 63 (2014) 1835 – 1841 1839

absolute difference in CO2 loading between the two temperatures at the fixed 15 kPa CO2 partial pressure. The
overall absorption enthalpy which describes the amount of heat involved during CO2 absorption was calculated
using the chemical speciation data and the Gibbs-Helmholtz relationship.8
While the conditions described here do not necessarily represent the optimum operating conditions it nevertheless
allows for a consistent comparison of the blends over a fixed range of conditions. It should also be noted that the
modelling approach described here is fundamentally based around the assumption that a combination of the
individual chemical models for the amines present in the solutions adequately describes the absorption chemistry
and no additional synergistic behaviours are required. Furthermore, the formation of insoluble precipitates is not
accounted for in the software and the solutions here are assumed to remain homogeneous. Thus, the predictions here
should be considered a guide only.

3. Results and Discussion

3.1 Overall CO2 mass transfer coefficients, KG

Overall CO2 mass transfer coefficients, KG, have been determined from the wetted wall flux data over a series of
CO2 loadings in each of the blends containing 3.0M BZA with 3.0M of the second amine. CO2 loading is expressed
here as the moles of CO2 per total mole of amine in the solutions. KG data is presented in Figure 1 together with
corresponding data for 6.0M MEA for comparison. The use of MEA solutions above 5.0M is uncommon
industrially in the absence of corrosion inhibitors; however, the data at 6.0M is fundamentally relevant here and
allows for a direct comparison of the BZA blends to MEA in the absence of additional effects stemming from
differences in amine concentration and CO2 loading.

4
Overall CO2 mass transfer co-efficient

6M MEA
3.5 3M BZA 3M MEA
3M BZA 3M DMEA
KG (mmol.m-2.s-1.kPa-1)

3 3M BZA 3M DEEA
3M BZA 3M AMP
2.5
3M BZA 3M 3-QUIN
2

1.5

0.5

0
0 0.1 0.2 0.3 0.4
CO2 loading (mole CO2/mole amine)

Figure 1. Plot of overall CO2 mass transfer coefficients, KG against CO2 loading at 40oC in blends containing 3.0 M BZA with
3.0M MEA, sterically hindered, and tertiary amines
1840 William Conway et al. / Energy Procedia 63 (2014) 1835 – 1841

The overall trend in KG at zero CO2 loading as a function of the second amine in the blend follows the trend MEA
~ DMEA ~ AMP > DEEA ~ 3-QUIN > standalone 6.0M MEA. A blend comprising BZA with MEA as components
consistently displayed the largest KG values due to the blend containing two fast reacting amines. Despite the
remaining blends containing only half the amount of reactive amine (the other half being sterically hindered or
tertiary amine) the majority of the blends were able to maintain faster or similar CO2 mass transfer than in
standalone MEA over the range of CO 2 loadings up to ~ 0.4. This observation relates to the individual roles of the
amines in the blends where the second amine is acting as a strong base to accept protons, while the concentration of
BZA is maximized for the fast reaction with CO2. The behavior is slightly reduced at higher CO2 loadings due to
increasing viscosity and the impact of this on diffusion processes. The larger and highly branched structure of
DMEA, DEEA, AMP, and 3-QUIN, account for the trend in KG as the CO2 loading of the solution increases
(compared to the blend containing BZA and MEA).

3.2 Vapour liquid equilibrium predictions

Results of estimating the absorption capacity, cyclic capacity and enthalpy of absorption for formulations with
different compositions are given in Table 4. Figure 5 shows vapour-liquid-equilibria data estimated at 40°C for the
formulations and Figure 6 the estimated enthalpy of absorption at 40°C as a function of CO 2 loading. Similar to the
KG data we have chosen here to compare the results to the corresponding data for 6.0M MEA. The mixtures show
comparable absorption capacities to each other and MEA. The result is not unexpected given the CO2 capacity of the
solutions are dictated here by the chemical pathway for CO2 via the stable BZA carbamate (and MEA carbamate in
the case of the BZA/MEA blend) and is largely indepedent of the second amine in the blend. Of the compositions
the largest cyclic capacity is achieved in a combination of BZA with AMP which is approximately double that
observed in 6.0M MEA. The magnitude of the absorption enthalpy in the blends is consistently larger than 6.0M
MEA across a broad range of CO2 loadings. It is this enthalpy that is responsible for the large cyclic capacity which
is strongly coupled to the protonation reaction of the amines. In the case of BZA, the significantly larger protonation
enthalpy, which is some 15 - 20kJ/mol larger than that of typical alkanolamines, is responsible for larger swings in
pH with variations in temperature, and thus accounts for larger swings in CO2 loading observed here.

Table 2. Absorption capacity, cyclic capacity, and overall absorption enthalpy in BZA blends
Blend Absorption Cyclic Overall absorption
capacity* capacity# enthalpy &
BZA/MEA 0.5 0.22 -85
BZA/AMP 0.51 0.23 -84
BZA/N,N-DMEA 0.49 0.21 -74
BZA/N,N-DEEA 0.51 0.17 -78
BZA/2,2-M-1-P 0.52 0.11 -77
BZA/3-QUIN 0.51 0.17 -79
6.0M MEA 0.51 0.13 -72

*
40oC, 15kPa CO2 #
40oC - 100oC, 15kPa CO2 &
40oC (kJ/mol)
 William Conway et al. / Energy Procedia 63 (2014) 1835 – 1841 1841

0
1.E+01

Absorption enthalpy (kJ/mole CO 2)

-20
CO2 partial pressure (Bar)

1.E-01

6M MEA -40
1.E-03
6M MEA
3M BZA 3M DEEA 3M BZA 3M DEEA
-60
1.E-05 3M BZA 3M DMEA 3M BZA 3M DMEA
3M BZA 3M 22M1P -80 3M BZA 3M 22M1P
1.E-07
3M BZA 3M 3-QUIN 3M BZA 3M 3-QUIN
3M BZA 3M MEA -100 3M BZA 3M MEA
1.E-09
3M BZA 3M AMP 3M BZA 3M AMP
1.E-11 -120
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

CO2 loading (mole/mole) CO2 loading (mole/mole)

Figure 2. (left) vapour-liquid-equilibria curves at 40°C estimated based on published equilibrium and thermodynamic data, horizontal line
corresponds to 15kPa CO2 partial pressure (right) differential absorption enthalpy as a function of CO2 loading at 40°C estimated based on
published equilibrium and thermodynamic data

Conclusions

CO2 absorption into aqueous amine blends containing BZA, MEA, and sterically hindered and tertiary amines
has been investigated here using a wetted-wall contactor and simulations of vapour liquid equilibrium using
published chemical equilibrium data. Based on the rapid CO2 mass transfer data and large absorption enthalpies and
cyclic capacities from the VLE simulations, several of the blends investigated here can be considered promising
candidates for further assessment as CO2 capture solvents.

Acknowledgements

This work was carried out within CSIRO’s Energy Flagship and it was supported by the Australian Government
through the Australia-China Joint Coordination Group on Clean Coal Technology (JCG). The views expressed
herein are not necessarily the views of the Commonwealth, and the Commonwealth does not accept responsibility
for any information or advice contained herein.

References
[1] Richner, G.; Carnal, A.; Conway, W.; Maeder, M.; Puxty, G. Thermokinetic properties and performance evaluation of benzylamine-based
solvents for CO2 capture. Chemical Engineering Journal 2014, Submitted for publication.
[2] Wei, C.-C.; Puxty, G.; Feron, P. Amino Acid Salts for CO2 Capture at Flue Gas Temperatures. Chemical Engineering Science 2014, 107, 218-
226.
[3] Puxty, G.; Rowland, R.; Attalla, M. Comparison of the Rate of CO2 Absorption into Aqueous Ammonia and Monoethanolamine. Chemical
Engineering Science 2010, 65, 915-922.
[4] Puxty, G.; Maeder, M. A Simple Chemical Model to Represent CO2-Amine-H2O Vapour-Liquid-Equilibrium. International Journal of
Greenhouse Gas Control 2013, 17, 215-224.
[5] Conway, W.; Bruggink, S.; Beyad, Y.; Luo, W.; Melian-Cabrera, I.; Puxty, G.; Feron, P. CO2 Absorption Into Aqueous Amine Blended
Solutions Containing Monoethanolamine (MEA), N,N-dimethylethanolamine (DMEA), N,N-diethylethanolamine (DEEA) and 2-amino-2-
methyl-1-propanol (AMP) for Post Combustion Capture Processes. Chemical Engineering Science 2014, Submitted for publication.
[6] Fernandes, D.; Conway, W.; Wang, X.; Burns, R.; Lawrance, G.; Maeder, M.; Puxty, G. Protonation Constants and Thermodynamic
Properties of Amines for Post Combustion Capture of CO2. Journal of Chemical Thermodynamics 2012, 51, 97-102.
[7] Fernandes, D.; Conway, W.; Burns, R.; Lawrance, G.; Maeder, M.; Puxty, G. Investigations of Primary and Secondary Amine Carbamate
Stability by 1H NMR Spectroscopy for Post Combustion Capture of Carbon Dioxide. Journal of Chemical Thermodynamics 2012, 54, 183-191.
[8] Kim, I.; Svendsen, H. F. Heat of Absorption of Carbon Dioxide (CO 2) i nMonoethanolamine (MEA) and 2-(Aminoethyl)ethanolamine
(AEEA) Solutions. Industrial and Engineering Chemistry Research 2007, 46, 5803-5809.